home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Cream of the Crop 20
/
Cream of the Crop 20 (Terry Blount) (1996).iso
/
bbs
/
wg960530.zip
/
CHANGES.DOC
next >
Wrap
Text File
|
1996-05-30
|
28KB
|
678 lines
W...W
W.W.W
.W.W.
WaterGate
Mailprocessor for Fidonet & Usenet
Changes and fixes to the previous version
(c) Copyright 1993-1996 WaterLine Software Development
All Rights Reserved
May 30th, 1996
Hello there!
It has been quite some time since the release of 0.90, but there
have been quite some things going one. No more but 12 different
patch files were release between these two major releases to test
new functionality and solve bugs.
I attempted to make the gateway completely transparent by putting
the Fido MSGID kludges in the Internet / Usenet Message-ID header
and the other way around and extracting and restoring them on the
way back. This way a proper In-reply-to header could be constructed
from the REPLY kludge, which was the original MSGID created on the
way in. This unfortunately didn't work.
A lot of tossers (including WaterGate in the beginning, <ahum>)
didn't like the new format of the MSGID kludge and were expecting
a valid Fido address as the first argument. I had to take out that
code again. Good news is that the code for putting Fidonet MSGID
and REPLY kludges in Internet / Usenet message headers is still
there though, so that part becomes transparent and allows linking
of replies, for one.
All these changes, testing, patches, awaiting feedback, fixing, etc.
took a lot of time, but the end result is that 0.91 is a stable
improvement over 0.90. See the new future.doc file for changes you
can expect in the near future.
Back to business..
Ramon
This file describes the changes and fixes that you will find in
version 0.91 of the WaterGate package. This version contains al lot
of bug fixes, partitial rewrites and some new features as listed
below.
■ Memory requirements have gone down drastically. WaterGate will
now used almost all the available "low" memory and won't write
messages to the "too big" directory as soon as it hits the 70k
limit anymore.
■ Points of nodes with a domain address are now automatically
mapping to username@p<point>.<user's domain address>. For
example, point 16 of node 2:200/111 with domain address
cold.wline.se will now get the e-mail address
user@p15.cold.wline.se and incoming mail for this address is
automatically mapped to the correct AKA.
■ Squish base code and the WtrUtil program are now a lot more
stable, have stopped destroying message bases all the time.
■ MSGID and REPLY kludges are now encoded in the Message-ID and
In-Reply-To: headers and decoded to the original format on the
way back.
■ Encoded files are now automatically decoded and stored in a
special directory. Currently only when importing a message in
a *.MSG base, but more functionality will follow as soon as
some problems have been solved (split part counting, for one).
■ Support for 25+ lines screens (windows) has been improved.
Centered menus, longer listings, not restoring to 25 lines
on exit, etc.
Following is the complete list of bugfixes and other improvements
between versions 0.90 and 0.91, in reverse order of fixing date.
PLEASE READ IT carefully (I know it is a lot) and see if something
relates to you. If it does, please check that it is still
functioning!
■ Now allowing continuation lines with MIME parameters when decoding.
Especially the NAME one.
■ Fixed some cleanup after Squish Purge, Link and re-index.
■ Removed the FD_Using log action.
■ Netmail on Hold is now processed when not running in FrontDoor mode.
■ Centered the copyright/beta info and about boxes.
■ Allowed quite some lists to use the full screen length instead of a
fixed length (groups list, areas list, users list, etc.)
■ Fixed some problems with writing the progress outside of the centered
percentage done windows on 25+ lines windows.
■ Video mode is no longer changed on exit if it wasn't set at startup.
This stop the 43 lines window from resetting to 25 lines. Neither
when crashing.
■ Fixed two potential memory freeing errors found via the P10 crashes.
■ Mail -> Netmail translation now uses the Gateway AKA in the MSGID
instead of the first system AKA.
■ Bad-case Message-ID -> MSGID conversion now builds the default MSGID
string instead of returning "0 00000000".
■ Fixed a small bug in the uu-decoding to do with the last line, but
only when it didn't contain a three-pair in the end.
■ Generated e-mail address for fido user will now contain "p0" when
the gateway aka is a point address as well. Without this information
it is impossible to map back to the node number (without the point).
Added for both normal and small address generation.
■ Small modifications the command-line help screens for WtrUtil and
WtrConf.
■ Took out some old UUCP routing code data structures and memory usage.
■ WtrUtil purge no longer calls the purging code when age and limit
are both set to 0.
■ WtrUtil Squish Purge code now closes the message base when done.
■ Attempted to fix the code that purges Squish and JAM areas by age
to prevent calculation errors on the maximum age of messages.
■ Removed "Not my outbound" in FrontDoor netmail checking code.
■ Added logging for news articles that are not accepted. The names of
all the newsgroups it was posted in are logged.
■ Bounce reason is no longer up in all uppercase, but the case from
the route.tdb file is kept.
■ Message-ID is no longer put into the MSGID header as this confuses
editors and tossers and causes loads of problems.
■ Message-ID in new style MSGID was interpreted as a Fido address to
form the complete From: aka, but completely failed. Built in some
extra checks to prevent this from happening.
■ Fixed a bug in the encoded file extracter for MIME encoded files. The
last or two last bytes could get lost at times.
■ Added prevention for bounced mail coming back to the list server after
a distribution by checking for "MAILER-DAEMON" and writing the message
to the bad area for inspection.
■ JAM base linking on disk should now work. A bug kept it from creating
a temporary file used to store the link information.
■ Squish index file is no longer opened and closed unnecessary if
there is nothing to update.
■ Last used Squish and JAM bases are now closed when WtrUtil Import
is done.
■ Squish base index is now updated on disk before the base is unlocked.
This should some problems in multi-tasker environments.
■ Read-ahead cache is now freed in advance when importing *.MSG msgs
into a JAM or Squish base.
■ When importing *.MSG messages into a JAM or Squish base, the
Soft-CRs are now removed.
■ New language entry for 1012: Disconnected from mailing list @1@.
■ Squish now closes the message base properly if an error occurs
during a scan.
■ Number of bytes exported from Squish base is now counted properly.
■ Scanned Squish messages are now properly flagged as Sent.
■ Added log file toggles plus help to WtrConf for the two new toggles.
■ Added NOLOCALFLAG option to ROUTE.TDB to remove the Loc flag before
importing a netmail into a *.MSG message base.
■ Message-ID: / In-Reply-To: contents are now properly kept during
transition in a FTN network and MSGID / REPLY are now properly kept
during transition on the Internet. The latter is the more important
though.
■ Built the MessageID -> MSGID coding and MSGID -> Message-ID decoding.
(wgmid$)
■ Built the MSGID -> Message-ID coding and Message-ID -> MSGID decoding.
(wgcid$)
■ Optional In-Reply-To: is now translated into a REPLY: kludge.
■ Message-ID: is now translated (coded, actually) into a MSGID: kludge.
Notice that translated mails never had MSGID up to now (oops...).
■ MSGID to Message-ID: and REPLY to In-Reply-To: or References: now
in place.
■ Started generic mapping between FTN MSGID: / REPLY: kludge and UUCP
Message-ID: / In-Reply-To: / References: